This repo contains a sparse learning library which allows you to wrap any PyTorch neural network with a sparse mask to emulate the training of sparse neural networks. It also contains the code to replicate our work Sparse Networks from Scratch: Faster Training without Losing Performance.
The library requires PyTorch v1.2. You can download it via anaconda or pip, see PyTorch/get-started for further information. For CUDA version < 9.2 you need to either compile from source, or install a new CUDA version along with a compatible video driver.
- Install PyTorch.
- Install other dependencies:
pip install -r requirements.txt
- Install the sparse learning library:
python setup.py install
MNIST and CIFAR-10 code can be found in the mnist_cifar
subfolder. You can run python main.py --data DATASET_NAME --model MODEL_NAME
to run a model on MNIST (--data mnist
) or CIFAR-10 (--data cifar
).
The following models can be specified with the --model
command out-of-the-box:
MNIST:
lenet5
lenet300-100
CIFAR-10:
alexnet-s
alexnet-b
vgg-c
vgg-d
vgg-like
wrn-28-2
wrn-22-8
wrn-16-8
wrn-16-10
Beyond standard parameters like batch-size and learning rate which usage can be seen by python main.py --help
the following sparse learning specific parameter are available:
--save-features Resumes a saved model and saves its feature data to
disk for plotting.
--bench Enables the benchmarking of layers and estimates
sparse speedups
--growth GROWTH Growth mode. Choose from: momentum, random, and
momentum_neuron.
--death DEATH Death mode / pruning mode. Choose from: magnitude,
SET, threshold.
--redistribution REDISTRIBUTION
Redistribution mode. Choose from: momentum, magnitude,
nonzeros, or none.
--death-rate DEATH_RATE
The pruning rate / death rate.
--density DENSITY The density of the overall sparse network.
--sparse Enable sparse mode. Default: True.
To run ImageNet with 16-bit you need to install Apex. For me it currently does not work to install apex from pip, but installing it from the repo works just fine.
The ImageNet code for sparse momentum can be found in the sub-folder imagenet
which contains two different ResNet-50 ImageNet models: A baseline that is used by Mostafa & Wang (2019) which reaches 74.9% accuravy with 100% weights and a tuned ResNet-50 version which is identical to the baseline but uses a warmup learning rate and label smoothing and reaches 77.0% accuracy with 100% weights. The tuned version builds on NVIDIA Deep Learning Examples: RN50v1.5 while the baseline builds on Intel/dynamic-reparameterization.
With the sparse learning library it is easy to run sparse momentum on your own model. All that you need to do is follow the following code template:
It is easy to extend the library with your own functions for growth, redistribution and pruning. See The Extension Tutorial for more information about how you can add your own functions.